Development and Validation of a Nigeria-Specific Preeclampsia Risk Prediction Model Incorporating Biochemical, Biophysical, and Socio-Contextual Markers

Existing first-trimester preeclampsia risk prediction models were developed in high-income country populations and have not been validated or adapted for the Nigerian obstetric population, where the epidemiology, risk factor profile, and biomarker distributions differ substantially. This dissertation develops and validates a Nigeria-specific preeclampsia risk prediction model incorporating biochemical, biophysical, and socio-contextual markers. A prospective multi-site cohort enrolled 800 women at 11 to 13 weeks gestation at teaching hospitals in Lagos, Kano, Enugu, and Port Harcourt, representing four geopolitical zones. Predictors assessed included mean arterial pressure, uterine artery pulsatility index by Doppler, placental growth factor (PlGF), pregnancy-associated plasma protein A (PAPP-A), and a novel socio-contextual risk composite (SCRC) incorporating food insecurity, night-time outdoor sleep exposure, and malaria burden. Preeclampsia developed in 11.4 percent by 34 weeks. Logistic regression and machine learning ensemble models were compared. The Nigeria-specific model incorporating SCRC achieved an AUC of 0.91 for preterm preeclampsia, compared to 0.79 for the Fetal Medicine Foundation model applied without local calibration. Cross-validation confirmed model stability. Decision curve analysis demonstrated clinical net benefit at threshold probabilities of 5 to 25 percent. The dissertation makes an original contribution to perinatal epidemiology by demonstrating that socio-contextual factors specific to Sub-Saharan Africa add independent predictive value to established biochemical-biophysical markers, supporting their integration into context-specific risk models. Keywords: preeclampsia, risk prediction model, Nigeria, first trimester screening, PlGF